Voice Gender Classification Using Combination of Gaussian Mixture Model and Support Vector Machine

Aji Kautsar Yusuf, Faizal Makhrus, S.Kom., M.Sc., Ph.D.

2024 | Skripsi | ILMU KOMPUTER

Klasifikasi gender berdasarkan audio adalah area penelitian penting yang telah mendapatkan perhatian signifikan dalam beberapa tahun terakhir. Kemampuan untuk secara akurat mengklasifikasikan gender individu berdasarkan sinyal audio mereka memiliki banyak aplikasi di berbagai bidang. Tujuan utama dari penelitian ini adalah membangun kombinasi GMM dan SVM untuk klasifikasi gender berdasarkan suara. Dua kombinasi dibangun dalam penelitian ini, Model Sekuensial dan Model Paralel. Kami mengevaluasi kinerja dari empat model yang berbeda: Model Sekuensial, Model Paralel, Gaussian Mixture Model (GMM), dan Support Vector Machine (SVM). Model Sekuensial menunjukkan akurasi tertinggi sebesar 79,11%, sementara Model Paralel mencapai akurasi sebesar 71,08%. GMM mencapai akurasi sebesar 72,67%, dan SVM mencapai 69,15%.

Gender classification based on audio is an important area of research that has gained significant attention in recent years. The ability to accurately classify the gender of individuals based on their audio signals has numerous applications in various fields. The primary objective of this research is to build a combination of GMM and SVM for gender classification based on voice. Two combination is built in this research, Sequential Model and Parallel Model. We evaluated the performance of four different models: Sequential Model, Parallel Model, Gaussian Mixture Model (GMM), and Support Vector Machine (SVM). The Sequential Model demonstrated the highest accuracy at 79.11%, while the Parallel Model achieved an accuracy of 71.08%. The GMM achieved an accuracy of 72.67%, and the SVM achieved 69.15%.

Kata Kunci : Voice Gender Classification, Voice Analysis, Machine Learning, Support Vector Machine, Gaussian Mixture Model, MFCC, Hybrid Model